Packing Protein Side-chain by Parallel Ant Colonies

Supplementary Information

Lijun Quan and Qiang Lü

  




Table 1. Score function and ACO parameters

PDB Score Function Score Terms α β
A B C D E F G H I J K L M N O P Q
1standard0.80.440.650.0040.490.561.171.171.171.10.525510031
2score120.80.440.650.0040.490.561.170.5851.171.1111110011
3score12\_full0.80.440.650.0040.490.561.170.5851.171.10.525510012
4score12minpack0.80.440.650.0040.490.561.170.5851.171111110013
5 score13 0.69210.17540.5253-0.007640.530.631.3220.33621.8830.525510.571021
7score130.69210.17540.5253-0.007640.530.631.3220.33621.8830.525510.571011
8pack_no_hb_env_dep0.80.10.650.0040.490.561.171.171.173.1111110131
6 RosettaHoles score The RosettaHoles scores are based on packing information about a cavity ball and the local region surrounding it, 1 2
{most importantly the contact surface area of atoms surrounding the cavity with respect to a sequence of probe radii.


Table 2. Score Terms

Score TermLabelDescription
fa_atrAlennard-jones attractive
fa_repBlennard-jones repulsive
fa_solClazaridis-jarplus solvation energy
fa_intra_repDlennard-jones repulsive between atoms in the same residue
fa_pairEpairwise electrostatics term derived from statistics on the pdb database
fa_dunFinternal energy of sidechain rotamers as derived from Dunbrack's statistics
hbond_lr_bbGlong range (beta or loop) backbone-backbone hydrogen bonds
hbond_sr\bbHshort range (helix) backbone-backbone hbonds
hbond_bb_sc Isidechain-backbone hydrogen bond energy
hbond_scJsidechain-sidechain hydrogen bond energy
dslf_ss_dstKdistance score in current disulfide
dslf_cs_angLcsangles score in current disulfide
dslf_ss_dihMdihedral score in current disulfide
dslf_ca_dihNca dihedral score in current disulfide
pro_closeOproline ring closure energy
envsmooth PStatistically derived fullatom environment potential
atom_pair_constraintQHarmonic constraints between atoms involved in Watson-Crick base pairs specified by the user in the params file


Fig. 1. Analysis of the performance of pacoPacker, CIS-RR and scwrl4 on different amino acids
(A) Percent correct within 40° for the χ1 angle. (B) Percent correct within 40° for the χ12 angle.
The test data used the 65 proteins set. Red, Green back slash and blue forward clash recolumns indicate
the test results of pacoPacker, CIS-RR and SCWRL4, respectively.



RMSD_CA RMSD_SA RMSD_SC
aCA Rmsd bCA Rmsd cCA Rmsd

Fig. 2. RMSD (Å) comparison on benchmark of CIS-RR



65CA_Chi1_40 65SA_Chi1_40 65SC_Chi1_40
aCA_χ1 bSA_χ1 cSC_χ1
65CA_Chi12_40 65SA_Chi12_40 65SC_Chi12_40
dCA_χ12 eSA_χ12 fSC_χ12

Fig. 3. Targets comparison on benchmark of CIS-RR (χ1 and χ12 within 40°)



65CA_Chi1_20 65SA_Chi1_20 65SC_Chi1_20
aCA_χ1 bSA_χ1 cSC_χ1
65CA_Chi12_20 65SA_Chi12_20 65SC_Chi12_20
dCA_χ12 eSA_χ12 fSC_χ12

Fig. 4. Targets comparison on benchmark of CIS-RR (χ1 and χ12 within 20°)



Table 3. Comparation of pacoPacker, CIS-RR and SCWRL4 on benchamrk of CIS-RR (65-protein)

Compare on benchmark of CIS-RR



Table 4. Analysis of the performance of pacoPacker, CIS-RR and scwrl4 on individual residues (65-protein test set)

PDB χ1(40°) χ12(40°)
SCWRL4 CIS-RR pacoPacker SCWRL4 CIS-RR pacoPacker
Axisscwrl4CISRRACOscwrl4CISRRACO
ARG79.8981.6683.0765.9668.9683.07
ASN84.9583.5186.5473.2072.3773.50
ASP86.4083.6888.0779.7974.6179.25
CYS87.1295.7692.800.000.000.00
GLN83.0581.6286.6065.6368.9761.00
GLU71.2876.1881.0857.0958.6160.14
HIS89.2489.2491.2979.5181.6081.18
ILE95.9096.4196.7585.6487.6980.00
LEU92.2493.0994.2985.8286.6780.56
LYS80.7981.5486.8262.9364.3065.15
MET85.3586.3688.8971.2174.7572.22
PHE96.0597.4398.0292.4994.2794.27
PRO87.8785.6288.4082.3080.3980.56
SER73.5071.3278.020.000.000.00
THR91.1390.8593.360.000.000.00
TRP91.7094.7691.7078.6086.4676.86
TYR94.6096.7795.4792.4493.3293.10
VAL90.6092.1992.320.000.000.00


Fig. 5. Analysis of the performance of pacoPacker, CIS-RR and scwrl4 on different amino acids
(A) Percent correct within 40° for the χ1 angle. (B) Percent correct within 40° for the χ12 angle.
The test data used the 377 proteins set (benchmark of SCWRL4). Red, Green back slash and blue forward
clash recolumns indicate the test results of pacoPacker, CIS-RR and SCWRL4, respectively. Note:
Because 2QOL cant be predicted by CIS-RR and 1G8Q miss mainchain atomsand, we exclude them from
the testing sets of 379 proteins used by the original paper of SCWRL4.



379RMSD_CA 379RMSD_SA 379RMSD_SC
aCA Rmsd bCA Rmsd cCA Rmsd

Fig. 6. RMSD (Å) comparison on benchmark of SCWRL4



379CA_Chi1_40 379SA_Chi1_40 379SC_Chi1_40
aCA_χ1 bSA_χ1 cSC_χ1
379CA_Chi12_40 379SA_Chi12_40 379SC_Chi12_40
dCA_χ12 eSA_χ12 fSC_χ12

Fig. 7. Targets comparison on benchmark of SCWRL4 (χ1 and χ12 within 40°)



379CA_Chi1_20 379SA_Chi1_20 379SC_Chi1_20
aCA_χ1 bSA_χ1 cSC_χ1
379CA_Chi12_20 379SA_Chi12_20 379SC_Chi12_20
dCA_χ12 eSA_χ12 fSC_χ12

Fig. 8. Targets comparison on benchmark of CIS-RR (χ1 and χ12 within 20°)



Table 5. Comparation of pacoPacker, CIS-RR and SCWRL4 on benchamrk of SCWRL4 (377-protein)

Compare on benchmark of SCWRL4



Table 6. Analysis of the performance of pacoPacker, CIS-RR and scwrl4 on individual residues (377-protein test set)

PDB χ1(40°) χ12(40°)
SCWRL4 CIS-RR pacoPacker SCWRL4 CIS-RR pacoPacker
Axisscwrl4CISRRACOscwrl4CISRRACO
ARG78.4679.3882.4165.7666.2082.41
ASN84.1182.9185.4474.2472.7575.76
ASP83.6481.9185.2374.9871.2475.21
CYS91.2093.3693.330.000.000.00
GLN79.0977.7881.5861.6960.8257.94
GLU72.9973.5578.5157.8856.3658.91
HIS88.5689.3190.1077.6779.4079.79
ILE95.8796.0695.7584.3584.9878.45
LEU92.1592.9993.3486.3386.7180.67
LYS77.3578.5383.1663.9365.7366.72
MET84.5385.1486.6371.0373.3471.62
PHE95.2695.7096.1892.7492.6792.51
PRO85.4285.1183.5981.7681.5878.02
SER71.5769.0975.980.000.000.00
THR90.7389.5891.960.000.000.00
TRP92.7994.0192.3480.4384.9277.03
TYR95.2994.9994.8592.7491.6191.70
VAL93.2393.4193.520.000.000.00


Table 7. Best conformations of pacoPacker are distributed on multiple threads

ID thread0 thread1 thread2 thread3 thread4 thread5 thread6 thread7
Quantity 29 31 35 29 33 39 107 139


minChi

Fig. 9. Comparisons between model1, model2 and model3 for χ1 and χ12 within 40°



minRMSD

Fig. 10. Comparisons between model1, model2 and model3 for RMSD



Table 8. Comparation of Model1 Model2 Model3 on a 50-protein test set selected from CIS-RR and SCWRL4

Comparation of Model1 Model2 Model3

 

Related documents batch download

Results Figure     Tables     Supplementary Information of PDB